kanerva, m.; korkiakoski, s.; lahtonen, k ... - aalto

This is an electronic reprint of the original article.This reprint may differ from the original in pagination and typographic detail.

Powered by TCPDF (www.tcpdf.org)

This material is protected by copyright and other intellectual property rights, and duplication or sale of all or part of any of the repository collections is not permitted, except that material may be duplicated by you for your research use or educational purposes in electronic or print form. You must obtain permission for any other use. Electronic or print copies may not be offered, whether for sale or otherwise to anyone who is not an authorised user.

Kanerva, M.; Korkiakoski, S.; Lahtonen, K.; Jokinen, J.; Sarlin, E.; Palola, S.; Iyer, A.;Laurikainen, P.; Liu, X. W.; Raappana, M.; Tervakangas, S.; Valden, MikaDLC-treated aramid-fibre composites

Published in:Composites Science and Technology

DOI:10.1016/j.compscitech.2018.11.043

Published: 08/02/2019

Document VersionPublisher's PDF, also known as Version of record

Published under the following license:CC BY-NC-ND

Please cite the original version:Kanerva, M., Korkiakoski, S., Lahtonen, K., Jokinen, J., Sarlin, E., Palola, S., Iyer, A., Laurikainen, P., Liu, X. W.,Raappana, M., Tervakangas, S., & Valden, M. (2019). DLC-treated aramid-fibre composites: Tailoringnanoscale-coating for macroscale performance. Composites Science and Technology, 171, 62-69.https://doi.org/10.1016/j.compscitech.2018.11.043

https://doi.org/10.1016/j.compscitech.2018.11.043

https://doi.org/10.1016/j.compscitech.2018.11.043

Contents lists available at ScienceDirect

Composites Science and Technology

journal homepage: www.elsevier.com/locate/compscitech

DLC-treated aramid-fibre composites: Tailoring nanoscale-coating formacroscale performanceM. Kanervaa,∗, S. Korkiakoskib, K. Lahtonenc, J. Jokinena, E. Sarlina, S. Palolaa, A. Iyerd,P. Laurikainena, X.W. Liud, M. Raappanac, S. Tervakangase, M. Valdenca Tampere University of Technology, Laboratory of Materials Science, PO Box 589, FI-33101, Tampere, FinlandbAalto University, School of Engineering, Department of Mechanical Engineering, PO Box 14300, FI-00076, Aalto, Finlandc Tampere University of Technology, Laboratory of Photonics, PO Box 692, FI-33101, Tampere, FinlanddAalto University, School of Chemical Engineering, Department of Chemistry and Materials Science, PO Box 14300, FI-00076, Aalto, Finlande DIARC-Technology Oy, Kattilalaaksontie 1, FI-02330, Espoo, Finland

A R T I C L E I N F O

Keywords:Aramid fibreInterfacial strengthPhotoelectron spectroscopy (XPS)Finite element analysis (FEA)Damage tolerance

A B S T R A C T

This work aims to quantify the effect of a diamond-like carbon coating (DLC) treatment of aramid fibres and toreveal the conversion of a fibre-level performance leap on the macroscale mechanical behaviour. The DLC-basedcoating is applied directly to the reinforcement and laminates are infused with an epoxy matrix. After char-acterisation of the coated surfaces, the performance of the composite is analysed via interlaminar shear testing,fatigue testing and damage tolerance testing, microbond tests, and 3D finite element simulation using a cohesivezone model of the interface. The results show that the coating treatment improves the fatigue life and the S-Ncurve slope for the laminates, while the residual strength after impact damage and environmental conditioning(water immersion at 60 °C) remains high. The scaling factor to convert the performance on macroscale wasdetermined to be 0.17–0.39 for the DLC-based fibre treatment.

1. Introduction

Aramid fibre reinforced plastics (AFRPs) are superior to carbon andglass fibre composites, not only in ballistic applications. AFRPs alsohave very low electrical conductivity, chemical inertness and highthermal resistance [1,2]. However, the interlaminar strength of AFRPstends to be low. The underlying cause of low interfacial strength is thelow adhesion of aramid fibres with essentially all traditional matrixpolymers and the weak skin of the fibres themselves [3,4]. The chal-lenge in modifying aramid fibre surfaces is that the chemical surfacetreatments, often being based on acidic solutions, damage the fibre bulkand lead to impaired mechanical behaviour [5,6]. Moreover, increasedadhesion between fibres and matrix does not necessarily lead to in-creased impact resistance but vice versa [7]. The solution to these issuesis to use an optimized, strong, and well adhering superficial layer. In-stead of high-priced test programmes with different length scales, aprocedure of microscale testing and modelling must be established todetermine the scaling of the performance improvement on a nanoscaletowards parameters on a macroscale, such as strength, fatigue life, anddamage tolerance.

One of the methods with the most potential for optimizing thecompatibility of aramid is the application of carbonous nanostructureson the fibre surface [8–10]. Diamond-like coatings (DLCs) in generalare tough, stiff and strong, and they adhere well on different substrates:recent studies have shown that optimised DLC compounds can alsointroduce durable bonding with carbon nanotube networks [11,12] aswell as adhesive polymers [13,14]. If the surface of aramid fibres couldbe strengthened by a strong coating and simultaneously improved interms of the matrix-fibre adhesion, the application range of thesecomposites will be widened enormously.

This work aims to enhance the potential of DLC-coated microscalefibres and laminate response. A coating treatment is applied to aramidfibre reinforcement and laminate structures are prepared using infusionof epoxy resin. The quality of the coating as well as the AFRP compositeis analysed, whereafter the mechanical performance of the laminates isstudied via interlaminar shear strength testing, fatigue testing, damagetolerance testing, microbond tests and finite element modelling. Theresults reveal the relation between the adhesion improvement and la-minates' quasi-static as well as dynamic, fatigue performance.

https://doi.org/10.1016/j.compscitech.2018.11.043Received 23 February 2018; Received in revised form 20 June 2018; Accepted 25 November 2018

∗ Corresponding author.E-mail address: [email protected] (M. Kanerva).

Composites Science and Technology 171 (2019) 62–69

Available online 30 November 20180266-3538/ © 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).

T

2. Materials and methods

2.1. Reinforcement, coating and lamination

We used commercial aramid-fibre (Twaron® 2200 1210 dtex, Teijin)reinforcement from a single batch and with a 2/2 twill weave, an arealweight of 173–175 g/m2 and with an areal bundle distribution of 50/50in the 0° and 90° direction, respectively. The reinforcement for the la-minates with a treatment was coated by DIARC-Technology Oy(Finland). A sheet of reinforcement was coated on one side at a time bycommercial carbon-based coating DIARC® Bindo. In the DIARC plasmacoating process, the part is treated in a vacuum at a low temperaturebelow 100 °C. Carbon ions with high kinetic energy form a thin (fromnanometres to micrometres) well adherent amorphous nanostructuredcoating when they hit the surface. Prior to the coating process, a seriesof the reinforcement samples were washed using a procedure, given inTable 1, to remove the processing finish (AT81).

The resin used for preparing AFRP composite laminates was a room-temperature-curing novolac epoxy resin (Araldite LY 5052, Aradur5052, Huntsman International) mixed using a hardener-resin ratio of38% (weight/weight). Laminates with a stacking sequence of [0°12] and[0°6] were manufactured on a mould by using the vacuum infusiontechnique (vacuum pressure 0.5 bar) for interlaminar shear strengthtesting and fatigue testing, respectively. The laminates for fatiguetesting (nominal thickness 1.5 mm) were cured in the mould after in-fusion under vacuum pressure of 0.3 bar for 24 h at room temperature.

2.2. Characterisation

2.2.1. Raman spectroscopyThe coating on the aramid filaments was studied using a Raman

spectroscope (LabRAM HR, Jobin Yvon Horiba) system using an Argonlaser ( = 488 nm; power= 4mW on sample) with a BX41 (Olympus)microscope and 50×/10× objective. Samples of reinforcement ( 1 cm1 cm) were analysed directly without further preparation. For data

comparison, linear background extraction was applied and curves werematched at 50 cm−1.

2.3. X-ray photoelectron spectroscopy (XPS) and X-ray Auger electronspectroscopy (XAES)

The DLC nature of the coating (sp2/sp3 ratio) was examined using anelectron spectrometer (Argus, Omicron Nanotechnology GmbH) utilisingnon-monochromatic Al Kα X-rays at 300W. Data was recorded at normalemission with a detection area of 2.93mm2. The XPS C 1s C–C/H andsp3 peaks were calibrated to 285.0 eV. The sp2/sp3 peak fitting wasperformed using component energy separation of 0.8 eV and properGaussian-Lorentzian line shapes of GL(90) tail(1.5) for sp2 andsymmetrical GL(80) for sp3. In addition, the XAES C KLL transition wasrecorded to analyse the sp2/sp3 ratio also from the D-parameter [15]defined by the energy separation between minima and maxima in adifferentiated C KLL spectrum. Samples were cut from the coating-treated and untreated reinforcement. To preserve the 0°/90° weave andneat fibre bundles given the relatively large analysis area, the sampleedges ( 1 cm 1 cm) were bonded using DP190 two-component epoxy(3M).

2.3.1. Contact angle measurementContact angle measurements were planned in order to estimate the

effect of the coating treatment on wettability. For initial assessment,5 μL droplets of purified water and diiodomethane (DIM) were placeddirectly on untreated (cleaned) and coating-treated aramid reinforce-ment. The device used was a model DSA100 (Krüss).

2.3.2. ImagingVisual light microscopy (DM 2500M, Leica) and scanning electron

microscopy (SEM) using a field-emission gun electron microscope(ULTRAplus, Zeiss) were used for studying the fibre surfaces andpolished laminate cross-sections. The samples after microbond testswere imaged directly without removing the samples off the microbondsample holder (after carbon evaporation). The nanoscale surfacemorphology was studied by using atomic force microscopy (AFM).Filament bundles were extracted from untreated and coating-treatedreinforcement fabric and attached on metal chips at two points. Theimaging was done by using a Dimension 3100 (Veeco/Bruker)apparatus in clean room conditions. Silicon tips (Nanosensors™) wereused and the measurements were carried out in a tapping mode at aconstant (0.21 Hz) scan rate. Height, amplitude, and phase-differencedata were collected. The post-processing of data was done by rotatingthe data to align fibre longitudinal axis, printing of transverse-fibreprofiles, and extraction of parabolic background to visualise thenanoscale details.

2.3.3. NanoindentationNanoindentation and nano-scratch were performed to study the

mechanical state and interfacial strength of the fibre-DLC coating in-terfaces on cross-sectioned AFRP laminate. A TI-900 TriboIndenter(Hysitron Inc.) testing system was used in the study. Based on the in-formation from SEM observation and trial-indentation, the measure-ments were conducted under a load-control mode with a constant forceof 200 μN and 50 μN for the indentation and scratch, respectively.Considering the rough cross-section surface and softness of the fibre, aspherical indenter (0.5 μm nominal tip radius) was selected to avoidexcessive ploughing and material pile-up.

2.4. Mechanical testing

2.4.1. Fibre tenacityThe possible effect of the coating on the aramid fibre filaments'

stiffness and strength was studied using a filament tester (Favigraph,Textechno) with 100 cN load cell and automatic filament gripping andtesting. A gauge length of 20mm and pre-tension of 0.60 cN/tex wasapplied. The tests were carried out at a 20mm/min displacement ratein a standard environment; 100 samples were analysed. Stresses werecalculated based on nominal (circular) cross-section confirmand viaSEM and nanoindentation.

2.4.2. Interlaminar shear strength testsInterlaminar shear strength (ILSS) tests and fatigue tests were per-

formed using a universal test machine (Dartec, 100 kN) with compu-terised control (Elite Suite, MTS). ILSS tests were performed for watercut specimens (3mm 8mm 20mm) in a three-point bend fixture(14mm span) according to ASTM D2344 (displacement rate 1.0 mm/min) for dried and conditioned (wet) specimens. The wet specimenswere immersed in distilled water (65 °C) for a week prior to testing; thedry specimens were dehydrated in a vacuum oven (50 °C, 0.5 bar) untilthey achieved constant weight [16]. For the wet specimens, threespecimens per series were tested. Average shear stress values werecalculated for six and three specimens for the dry and wet series, re-spectively.

2.4.3. Fatigue testsThe laminates were post-cured in an oven for 15 h at 60 °C and

Table 1Procedure for washing reinforcement (sizing) prior to the coating treatment.

Step Action Medium

1 Ultrasonic cleaning (50 °C, 10min) Tap water2 Ultrasonic cleaning (30 °C, 10min) Ethanol3 Ultrasonic rinsing (50 °C, 5min) Purified water (Millipore Elix 10)4 Dehydration in vacuum oven (70 °C, 14 h,

0.5 bar)Air

M. Kanerva et al. Composites Science and Technology 171 (2019) 62–69

63

water-cut, dog-bone shaped [17] specimens were prepared for the fa-tigue tests. The specimens were provided with glass fibre-reinforcedepoxy (GFRE) tabs, bonded adhesively as described in previous studies[17]. Initial tensile moduli were measured prior to fatigue loading forall the specimens using a 50mm extensometer and 7 kN/min load rate(stress-strain curve fit over 0.001–0.003m/m range). A constant-loadamplitude, sinusoidal wave-form uniaxial tensile loading with a stressratio of R=0.1 was applied for fatigue testing. A test frequency of 4 Hzwas used to avoid excessive accumulation of heat. The fatigue test re-sults were analysed as stress-number of cycles (S-N) graphs and therelationship between the maximum stress and the number of cycles tofailure was formed using a power-law regression =N C Sm; the thick-ness and width measured at the specimen's gauge section, at threedifferent points, were used to calculate averaged cross-sectional areaand further the engineering stress (S). Here, the least-squares methodwas used according to ASTM E 739 to obtain parameters C and m; thenumber of cycles was used as a dependent variable.

2.4.4. Damage tolerance assessmentThe damage tolerance assessment aimed to reveal the resistance of

the laminates with coating-treated and untreated reinforcement againstimpact loads and subsequent ageing in a harsh environment.Rectangular specimens (27mm × 170mm, gauge length 110mm) werewater-cut from the laminate used for the fatigue test specimens. Thespecimens were provided with adhesively bonded GFRE tabs, in a si-milar way to the fatigue specimens. Two series (coating-treated/un-treated) of impact-damaged specimens were produced by subjecting animpact (0.5 J–8.9 J) per specimen using impact apparatus according toASTM D 5628-10 and a 15.9mm spherical impactor. A carbon fibre-reinforced epoxy backing plate with a 25mm hole for impact point andmechanical clamps were applied to give proper support to the spe-cimen. After the impact loading, the specimens were immersed in dis-tilled water and placed in an environmental chamber (60 °C) for fourweeks. Finally, the specimens were tested using a universal tester (seeSection 2.4.2) at a 7 kN/min loading rate and in wet conditions (eachspecimen removed from water immersion within two minutes beforethe test start). The residual strength of each specimen was determinedbased on the intact (net) cross-section and peak load. The quality of theintact specimens as well as the damage size per specimen after impactwere determined using ultrasonic inspection prior to impact loading,after impact loading and after the water immersion. Ultrasonic in-spection device (Omniscan MX, Olympus) was used in a transmission(2.25MHz) mode and inspection operated under water.

2.4.5. Microbond testsThe adhesion between the matrix and single fibres was studied using

a FIBRObond micro-droplet tester [18]. Here, the tester was set to use a

1 N load cell, surgeon knives (R35, thickness 0.254mm, Feather, Japan)for droplet loading, and an electrical linear motor to load the droplet.The droplets were prepared using a developed resin dip method withthe same resin batch as for the laminates (Section 2.1). Coating-treatedfilaments were extracted from the outer surface of the reinforcementbundles. The droplet fibre samples were cured in an oven (50 °C) for18 h. Droplets of different sizes were prepared to observe the effect ofdroplet size on the ultimate, peak force. The tests were performed at aconstant displacement rate until detachment of the droplet, and thepeak force and embedded area were used to calculate interfacialaverage shear strength (IFSS).

2.5. Interface modelling

The interface between the fibres and the matrix was modelled on afinite element (FE) basis using Abaqus® 2017, standard (DassaultSystèmes). The target was to form a mechanical model of the interfacethat covers the elastic behavior and damage. The 3D geometry of un-treated and coating-treated fibre droplet samples were modelled basedon their microscopy images (Fig. 1) and the rest of the test setup wasmodelled to account for the overall deformation. The sample holderwas meshed using quadratic elements (C3D10) and the rest of thesystem using linear tetras (C3D4), wedges (C3D6) and bricks (C3D8R).The sample holder was modelled as linear cast acrylic (Young's modulus3.3 GPa, Poisson's ratio 0.35), the surgeon knives as linear steel (mod-ulus 210 GPa, Poisson's ratio 0.3), fibre as linear aramid (Young'smodulus 108 GPa, Poisson's ratio 0.3), and the resin droplet as initiallylinear epoxy following ideal plastic yield (Young's modulus 3.2 GPa,Poisson's ratio 0.35, yield stress 84MPa, ultimate strength 86MPa at0.09m/m plastic strain). Residual stresses due to the cure at elevatedtemperature were included applying a linear thermal expansion of 71μm/(m K) and 37 μm/(m K) for the epoxy and aramid, respectively, andthermal loading of T=−30 °C (cool-down), by using a separate stepin the finite element analysis (FEA).

The fibre matrix interface was modelled using cohesive zone (CZM)elements with a bilinear traction-separation law, initial stiffness of1015 N/m3 for all three fracture modes, and the following damage onsetcriterion:

=max , , 1 ,ap

c

ap

c

ap

c

1, 2, 3,

(1)

where τ is traction, sub-indices 1, 2, 3 refer to the fracture modes (modeI, mode II, mode III, respectively), ap to a momentary value per appliedload level, and c to a critical value. Interfacial damage evolution wasgiven a critical energy release rate limit,Gcr , to define full debond at theinterface. The crack-tip mode mixity upon crack propagation was takeninto account using a linear interaction law:

Fig. 1. FE modelling of fibre droplet samples and the test setup: (a) droplet modelling based on a microscopy image; (b) the FE model of the entire micro-droplet testsetup.


64

= + +f GG

GG

GG

,I

cr

II

cr

III

cr (2)

where G is energy release rate, and sub-indices I, II, III refer to thefracture modes.

3. Results

3.1. Fibre and coating characterisation

The SEM imaging indicated that the DLC coating-treated fibre sur-face was smooth and was well wetted by the resin in the laminate(Fig. 2(a)). The fibre testing resulted in an average ultimate load of0.313 ± 0.094 N and 0.320 ± 0.096 N (i.e. 3370–3290MPa) for theuntreated and coating-treated fibres, respectively (full curves inFig. 2(b)). Clearly, the coating resembled nanoscale surface treatmentand did not affect the bulk fibre behavior significantly. Overall, theultimate stress level is of the order of high-strength aramid fibres, such

as Kevlar 49® (DuPont). The SPM (scanning probe microscopy) imagingand nanoindentation data confirmed a filament diameter of 11 μm andalso clean, well-wetted interfaces in the epoxy-infused laminates, asshown in Fig. 4(a). The Raman spectra for coating-treated and un-treated filaments are shown in Fig. 4(b), where the reference spectrashow typical output for aramid surface [19,20]. Analysis of the DLCcharacter on the coating-treated fibres was deemed unreliable due tothe DLC hump [12,21] overlapping with the aramid peaks.

The surface morphology of untreated and coating-treated single fi-laments was studied by using AFM (see Fig. 3). The vacuum depositionmethod applied for the coating, resembling a kind of hybrid betweenchemical and physical vapour deposition, forms a slightly more roughsurface compared to the reference, untreated fibres. It should be notedthat, due to coating fabric instead of fibres, the coating thickness, aswell as the surface structure vary locally. Therefore, the imaging, suchas in Fig. 3(b), represents only local, nanoscale roughness on the sur-face.

Fig. 2. Fibre characterisation: (a) force-extension curves; (b) SEM of untreated and coating-treated fibres, and cross-sections of the fibres in laminates.

Fig. 3. AFM analysis of fibres: (a) untreated filament surface; (b) coating-treated filament surface. For all data shown, the fibre longitudinal axis runs approximatelyin the horizontal direction.


65

The DLC coating on the aramid filaments was further analysedbased on XPS/XAES data of the coating-treated reinforcement surface.The C 1s peak and the 1st derivative of the KLL region HR spectra areshown in Fig. 5. Pure diamond (ta-C type) coatings are based on the sp3

carbon bond whereas graphite represents pure sp2 bonding (a:C type).Here, the analysis of the two synthetic components of the C 1s fityielded a semi-quantitative [21] sp2 content of 44–56% for the coatingsurface. The interpretation of the D-parameter in turn supported a sp2

content of 40–53% verifying that the DLC coating had a significant sp3

fraction. In general, the properties of DLC coatings can be adjusted byprocess parameters, and the sp3/sp2-ratio optimised per application;

values of 50–100% being typical [21].The surface wettability for polar and nonpolar liquids (water, DIM)

was observed to be very good. It was not possible to measure dropletcontact angles due to the droplets were instantly absorbed into thereinforcement (Fig. 6)—suggesting a high surface energy level for un-treated and coating-treated fibers.

3.2. Mechanical performance of laminates

The ILSS values for the laminates with either coating-treated oruntreated reinforcement are shown in Fig. 7(a). The washing process

Fig. 4. Coating characterisation: (a) nanoindentation on coating-treated reinforcement; (b) typical Raman spectra of untreated and coating-treated reinforcement.

Fig. 5. XPS/XAES analysis of the coating-treated reinforcement surface from two distinct measurement positions showing typical spectra and variation observed onthe surface: (a) C 1s peak and fit; (b) derivative of the C KLL spectrum and determination of the D-parameter value.

Fig. 6. Contact angle study directly on fabrics: (a) untreated (washed) fabric; (b) coating-treated fabric.


66

significantly affected the coating-fibre compatibility—washing in-creased the ILSS value by 86% (dry state) for the coating-treated la-minates, whereas the effect of washing was negligible in the untreatedcase. The average ILSS for the coating-treated samples with the washingprocess was found to be 4.0–4.2% higher compared to the fully un-treated samples, depending on the test conditions (dry/wet). The ILSSvalues for both the reference laminates and the coating-treated lami-nates with the washing process (Table 1) were 20–43% higher thanreported values in the current literature for H3PO4 and brominationtreated aramid-epoxy laminates [6,22].

Fatigue of composites in general is sensitive to fibre surface treat-ments and can reveal any underlying potential to strengthen aramid-epoxy interfaces. The fatigue test results are shown in Fig. 7(b) and itcan be seen that the fatigue lives significantly improved throughout theapplied stress range due to the coating treatment. The slope parameterof the S-N curve fit resulted in values of −25.47 and −27.80 for theuntreated and coating-treated laminates, respectively. The initialYoung's moduli indicated slightly higher values for the coating-treatedspecimens (29.72 ± 0.9 GPa) compared to un-treated specimens(28.03 ± 0.5 GPa)—yet the difference is within experimental scatterdue to measured laminate thickness variation. The ultrasonic inspectionresults (Fig. 8(a)) for the damage tolerance analysis revealed greaterdamage per impact energy for the laminates with coating-treatedreinforcement. However, based on the damage tolerance tests, theresidual strength of the coating-treated laminate specimens after impactand water exposure remained the same or was higher (+9.4 5%,0–2.2 J impacts) throughout the impact energy range compared to theuntreated specimens (Fig. 8(b)). In the current literature, improved in-plane mechanical properties of aramid-epoxy laminates due to any fibretreatment have led to impaired impact properties [7,22].

3.3. Interface strength and model parameters

The results of the microbond tests are shown in Fig. 9(a). Based onthe peak forces and individual embedded areas of droplets, statisticalIFSS values of 17.6 3.6MPa and 21.7 5.2MPa were calculatedfor the untreated and coating-treated samples. The results indicateimprovement in adhesion ( 24%) with the coating treatment. FEA wascarried out to validate the experimental IFSS values; the fitted responsecompared to raw force-displacement data is shown in Fig. 9(b). Thefitting resulted in model parameters of c =22.7MPa and Gcr =650 J/m2 for the untreated filament droplet sample, and c =22.2MPa andGcr =500 J/m2 for the coating-treated filament droplet sample (re-spective IFSS values 21.7MPa and 20.3MPa). These results establishthe validity of microbond IFSS values to be used in the determination ofthe onset criterion for the developed interface model. The fitted tractionvalues are realistic when compared to the microbond tests and reportedinterfacial stress values related to failure—although direct comparisonwith the reported simulations cannot be made due to the simplifiednon-3D (axisymmetric) models in the current literature, e.g. by Nishi-kawa et al. [23] and Sato et al. [24].

The samples were imaged after testing, and typical fracture surfacesare shown in Fig. 10. The imaging reveals a skin-like layer on thefracture surface of the coating-treated samples and it might resemblethe DLC coating. Skin-like, torn strips are also shown on the fracturesurfaces of untreated samples but they can be shredded bulk filamentsurface. In general, based on AFM and SEM images, the surfaces of theuntreated fibres often contain grooves and torn off-pieces. Our hy-pothesis is that the coating might form a protective layer preventingfracture of the bulk fibre surface. Additionally, the sharp-edged grooveson untreated filaments might work as stress concentration points and

Fig. 7. Mechanical testing results: (a) average ILSS results (R= untreated series, D= coating-treated series); (b) fatigue data and determined S-N curves.

Fig. 8. Damage tolerance analysis: (a) ultrasonic inspection of the test specimens before and after impact and conditioning (Ri=untreated series, Di=coating-treated series); (b) residual strength after impact and conditioning.


67

lead to premature interfacial failure on untreated filaments.

4. Discussion

The direct effect of a surface treatment for fibres is defined by theinterphase formed when embedded by a matrix. Here, a 3D crackpropagation model was fitted for fibre-matrix interfaces based on force-displacement data from microbond tests. The determined traction limit

c of damage initiation at the untreated and coating-treated interfacesverified the applicability of bare IFSS values —typically derived basedon pure microbond test data. The modification of aramid fibre interfacemust be optimised per application, and the surface treatment can evi-dently be adjusted using the microbond test. Furthermore, a fitted in-terface model can be applied to describe the interfacial performance forlarger models, representative volume elements (RVEs), and sub-modelsfor structural design [24,25]. Thus, for industrial applications, optimi-sation of the DLC coating for aramid fibre reinforcements can be carriedout faster and at markedly lower costs using the microbond-FEA pro-cedure prior to coupon and certainly before component-level testing.Due to the nature of microbond tests [26,27], hundreds of tests areneeded to achieve reliable statistical significance, which formerly hasnot been possible. Finally, what is essential is how much the adhesion atfibre matrix interfaces must be improved to gain a certain enhancementon the macroscopic performance, such as strength and fatigue life. Inthis study, the IFSS values supported by FEA suggested an adhesionenhancement of 24% on the fibre level. Scaling factor values of 0.17,0.38 and 0.39 must be then applied to estimate the observed im-provement in ILSS (4% improvement), S-N curve slope (9.14% im-provement) or residual strength after low energy impact and

conditioning (9.4% improvement), respectively. The scaling factorgives a convenient indication of the transfer of the benefits of fibre-matrix adhesion and it clearly establishes the significance of the na-noscale surface treatment on aramid fibres.

The mechanism of interfacial strengthening by the DLC coating wasconcluded to occur via protection or strengthening of the bulk fibresurface, in addition to strong adhesion to the epoxy matrix; the coating-matrix interface was assessed to be the point of failure initiation duringthe microbond tests. Similar defect-covering characteristics for surfacetreatments of carbon-fibres have been reported in the current literature[28,29].

5. Conclusions

In this study, aramid-fibre fabric was directly treated on its outersurfaces using a DLC-based coating and composite laminates wereprepared with epoxy resin using vacuum infusion. The experimentalresults of the characterisation and laminate testing confirmed the fol-lowing: in the case of coating treatment, the ultrasonic washing processof the aramid reinforcement improved the interlaminar shear strength(ILSS) by 86%. The ILSS of the laminate with the wash and coatingtreatment on the reinforcement was comparable or better ( 4%) com-pared to the untreated laminate with off-the-shelf reinforcement in bothdry and wet conditions. For the coating-treated laminates, the slopeparameter of the S-N curve fitted to the tensile-tensile fatigue dataimproved by 9.14% compared to the results of the untreated laminate.Also, the residual static strength after conditioning (water immersion at60 °C) and moderate impact (0–2.2 J) was on average 9.4 ± 5.0%higher for the coating-treated laminate compared to untreated

Fig. 9. Microbond test results and interface simulation: (a) test results; (b) representative FEA results for untreated (droplet 12) and coating-treated (droplet 40) tests.

Fig. 10. Fracture surfaces after microbond tests: (a) untreated filament-droplet interface; (b) coating-treated filament-droplet interface.


68

laminate, suggesting improved damage tolerance. The results show thatan optimised surface treatment can improve the overall performance ofaramid fibre reinforced structures. Moreover, the finite element simu-lations and microbond tests revealed that the scaling of the filament-level adhesion improvement is less than unity to reach the macroscaleeffect: in the case of the DLC-based coating and the aramid-epoxy la-minate, the scaling factor was 0.17–0.39 depending on the parameterson a macroscale (ILSS, S-N curve slope, residual strength after impact).

Acknowledgements

This investigation was funded by a grant from the Finnish Metalsand Engineering Competence Cluster (FIMECC HYBRIDS) and by theHISCON (259595) project from Academy of Finland. Authors ac-knowledge Kevra Oy and researchers H. Niemelä-Anttonen and M.Kakkonen for assistance with experimental activities.

References

[1] S. Luo, W. Van Ooij, Surface modification of textile fibers for improvement of ad-hesion to polymeric matrices: a review, J. Adhes. Sci. Technol. 16 (13) (2002)1715–1735.

[2] K. Imielińska, L. Guillaumat, The effect of water immersion ageing on low-velocityimpact behaviour of woven aramid–glass fibre/epoxy composites, Compos. Sci.Technol. 64 (2004) 2271–2278.

[3] P. de Lange, P. Akker, E. Mäder, S.L. Gao, W. Prasithphol, R. Young, Controlledinterfacial adhesion of Twaron aramid fibres in composites by the finish formula-tion, Compos. Sci. Technol. 67 (10) (2007) 2027–2035.

[4] P. Tarantili, A. Andrepoulos, Mechanical properties of epoxies reinforced withchloride-treated aramid fibers, J. Appl. Polym. Sci. 65 (1997) 267–276.

[5] T. Lin, S. Wu, J. Lai, S. Shyu, The effect of chemical treatment on reinforcement/matrix interaction in Kevlar-fiber/bismaleimide composites, Compos. Sci. Technol.60 (2000) 1873–1878.

[6] J. Lin, Effect of surface modification by bromination and metalation on Kevlar fibre-epoxy adhesion, Eur. Polym. J. 38 (2002) 79–86.

[7] M. Shaker, I. Kamel, F. Ko, J. Song, Improvement of the interfacial adhesion be-tween Kevlar fiber and resin by using R-F plasma, J. Compos. Technol. Res. 18 (4)(1996) 249–255.

[8] D. Devlin, K. Hubbart, K. Siebein, T. Archuleta, D. Coates, Diamond like carboncoated Kevlar for improved mechanical properties, 13th ICCVD Conference; vol. 96.Electrochemical Society Proceedings, US, 1996, pp. 691–698.

[9] Devlin, D., Coates, D., Archuleta, T.. Diamond-like-carbon coated aramid fibershaving improved mechanical properties. 2002. US Patent 6,432,537.

[10] S. Palola, E. Sarlin, S. Kolahgar Azari, V. Koutsos, J. Vuorinen, Microwave inducedhierarchical nanostructures on aramid fibers and their influence on adhesionproperties in a rubber matrix, Appl. Surf. Sci. 410 (2017) 145–153.

[11] H. Schittenhelm, D. Geohegan, G. Jellison, A. Puretzky, M. Lance, P. Britt, Synthesisand characterization of single-wall carbon nanotube-amorphous diamond thin-film

composites, Appl. Phys. Lett. 81 (2002) 11.[12] A. Iyer, A. Kaskela, L.S. Johansson, L. Xuwen, E. Kauppinen, J. Koskinen, Single

walled carbon nanotube network—tetrahedral amorphous carbon composite film,J. Appl. Phys. 117 (22) (2015) 225302.

[13] J. Takahashi, A. Hotta, Adhesion enhancement of polyolefins by diamond likecarbon coating and photografting polymerization, Diam. Relat. Mater. 26 (2012)55–59.

[14] J. Aakkula, J. Jokinen, O. Saarela, S. Tervakangas, Testing and modelling of Diarcplasma coated elastic plastic steel wedge specimens, Int. J. Adhesion Adhes. 68(2016) 219–228.

[15] A. Mezzi, S. Kaciulis, Surface investigation of carbon films: from diamond to gra-phite, Surf. Interface Anal. 42 (2010) 1082–1084.

[16] J. Posio, Interlaminar Shear Strength Testing of Non-unidirectionally ReinforcedComposites, M.Sc. thesis Aalto University, Finland, 2016.

[17] S. Korkiakoski, P. Brondsted, E. Sarlin, O. Saarela, Influence of specimen type andreinforcement on measured tension–tension fatigue life of unidirectional GFRP la-minates, Int. J. Fatig. 85 (2016) 114–129.

[18] M. von Essen, E. Sarlin, O. Tanhuanpää, M. Kakkonen, P. Laurikainen,M. Hoikkanen, et al., Automated high-throughput microbond tester for interfacialshear strength studies, The SAMPE Europe Conference. Stuttgart, Germany,2017,November, pp. 14–16.

[19] L. Penn, F. Milanovich, Raman spectroscopy of Kevlar 49 fibre, Polymer 20 (1)(1979) 31–36.

[20] H. Cen, Y. Kang, Z. Lei, Q. Qin, W. Qiu, Micromechanics analysis of Kevlar-29aramid fiber and epoxy resin microdroplet composite by Micro-Raman spectro-scopy, Compos. Struct. 75 (1–4) (2006) 532–538.

[21] X. Yan, T. Xu, S. Yang, H. Liu, Q. Xue, Characterization of hydrogenated diamond-like carbon films electrochemically deposited on a silicon substrate, J. Phys. D Appl.Phys. 37 (17) (2004) 2416–2424.

[22] S.J. Park, M.K. Seo, T.J. Ma, D.R. Lee, Effect of chemical treatment of Kevlar fiberson mechanical interfacial properties of composites, J. Colloid Interface Sci. 252(2002) 249–255.

[23] M. Nishikawa, T. Okabe, K. Hemmi, N. Takeda, Micromechanical modeling of themicrobond test to quantify the interfacial properties of fiber-reinforced composites,Int. J. Solid Struct. 45 (14) (2008) 4098–4113.

[24] M. Sato, E. Imai, J. Koyanagi, Y. Ishida, T. Ogasawara, Evaluation of the interfacialstrength of carbon-fiber-reinforced temperature-resistant polymer composites bythe micro-droplet test, Adv. Compos. Mater. 26 (5) (2017) 465–476.

[25] A. Hodzic, S. Kalyanasundaram, A. Lowe, Z. Stachurski, The microdroplet test:experimental and finite element analysis of the dependance of failure mode ondroplet shape, Compos. Interfac. 6 (4) (1998) 375–389.

[26] Y. Hou, T. Sun, An improved method to make the microdroplet single fiber com-posite specimen for determining the interfacial shear strength, J. Mater. Sci. 47 (1)(2012) 4775–4778.

[27] M. Pitkethly, J. Favre, U. Gaur, J. Jakubowski, S. Mudrich, D. Caldwell, et al., Around-robin programme on interfacial test methods, Compos. Sci. Technol. 48 (1)(1993) 205–214.

[28] X. Zhang, X. Fan, C. Yan, L. Hongzhou, Y. Zhu, X. Li, et al., Interfacial micro-structure and properties of carbon fiber composites modified with graphene oxide,ACS Appl. Mater. Interfaces 4 (1) (2012) 1543–1552.

[29] J. Dong, C. Jia, M. Wang, X. Fang, H. Wei, H. Xie, et al., Improved mechanicalproperties of carbon fiber-reinforced epoxy composites by growing carbon black oncarbon fiber surface, Compos. Sci. Technol. 149 (1) (2017) 75–80.


69

kanerva, m.; korkiakoski, s.; lahtonen, k ... - aalto

Documents